Interaction analysis

Hi,

I’m trying out the interaction analysis module of Disvis, but I get the following error:

  ValueError: Some selected residues where not found in the PDB file. Please check your input residues.

My scanning chain pdb is composed only of HETATMs. Could that be the cause? Is there a workaround?
Thanks,

Marta

Simply try changing the atom type to ATOM instead of HETATM

My scanning chain is a non-standard residue, if I use atom type as ATOM I get right away a message saying the PDB file is wrongly formatted.

Be careful not to mess up the formatting of the PBD file when replacing HETATM by ATOM
You have to include two spaces when replacing HETATM by ATOM

Thank you for the help previously. It worked once I changed the atom type to ATOM and renamed one of the atoms to “CA”.

I am now repeating some runs, and I’m getting very different results in the interaction analysis: Earlier this month I got clear results, with ~ 1/3 of my fixed chain selected residues showing a higher frequency than the others. When I repeat the run now (same parameters, same input files), all my selected residues have the same frequency.

Has something changed in the way the interaction analysis is performed?

I am now repeating some runs, and I’m getting very different results in the interaction analysis: Earlier this month I got clear results, with ~ 1/3 of my fixed chain selected residues showing a higher frequency than the others. When I repeat the run now (same parameters, same input files), all my selected residues have the same frequency.

Would you mind to provide the IDs of the runs that present different results (by email or private message)? Did you run them on different resources (CPU vs GPU)?

Thank you for the help previously. It worked once I changed the atom type to ATOM and renamed one of the atoms to “CA”.

For the record, here is the email answer to the failing interaction analysis when analysing ligands:

Unfortunately, in its current version, DisVis analyses interaction by looking for “CA” atoms for each residue listed in the interaction file. This will be changed in the next version (DisVis 3.0) that we will release soon.
In the meantime, one workaround would be to rename one of the ligand atom to “CA” in order to bypass the issue.
We do not exclude that other input parameters might be fine tuned to deal with ligands, we mainly optimised DisVis for protein-protein cases.
Reducing the “Minimum interaction volume” significantly might help you to get better results for instance.

Thanks for your feedback,

Cordially,

Mikael

Hi Mikael,

I’ve sent you the run IDs some days ago by message. Did you have a chance to take a look at it?

Thanks,
Marta

Hello Marta,

We are currently investigating the issue, we “closed” the grid version for the moment. Please refer to the “local” results, we will contact you as soon as we have identified and solved the problem. As you might have noticed, this impacts only the interaction analysis results, you can trust the rest of the output for both versions.

Sorry for the delay in my answer, it took more time than expected,

Mikael

Hello Marta,

After some further testing, it seems that the error between the CPU and GPU versions shows up when the “maximum clash volume” and “minimum interaction volume” are too high compared to the ligand volume. I can send you by message some identical runs (CPU and GPU) we have with a reduced volume for both parameters using your input data (PDBs, restraints and selection), we used half of the ligand volume in those runs.

We will try to fix this issue in the next version of DisVis, in the meantime I would advise to carefully check the parameters when using a ligand, we are still not officially supporting them.

We’re again sorry for the time it took to figure out the issue,

Mikael

Hi,

I had the same error:
ValueError: Some selected residues where not found in the PDB file. Please check your input residues.

For some reason I had double occupancies in one of the .pdb files. E.g. for the C-alpha atom of Met137:

ATOM 1165 N MET F 137 -51.163 5.260 -23.047 1.00 55.94 N
ATOM 1166 CA AMET F 137 -51.070 6.659 -23.479 0.55 59.87 C
ATOM 1167 CA BMET F 137 -51.034 6.654 -23.468 0.45 59.87 C
ATOM 1168 C MET F 137 -51.180 7.610 -22.289 1.00 60.49 C
ATOM 1169 O MET F 137 -50.426 8.573 -22.173 1.00 61.12 O
ATOM 1170 CB AMET F 137 -52.158 7.008 -24.510 0.55 60.96 C
ATOM 1171 CB BMET F 137 -52.053 7.004 -24.557 0.45 60.95 C
ATOM 1172 CG AMET F 137 -52.133 6.227 -25.838 0.55 62.40 C
ATOM 1173 CG BMET F 137 -51.672 6.516 -25.959 0.45 63.84 C
ATOM 1174 SD AMET F 137 -50.694 6.458 -26.932 0.55 72.63 S
ATOM 1175 SD BMET F 137 -50.063 7.101 -26.583 0.45 72.11 S
ATOM 1176 CE AMET F 137 -51.194 5.535 -28.387 0.55 48.25 C
ATOM 1177 CE BMET F 137 -50.337 8.865 -26.716 0.45 56.47 C

The solution was to delete double occupancies in my .pdb file:
(E.g. Delete the line 1167 and the “A” of “AMET” in line 1166)

ATOM 1165 N MET F 137 -51.163 5.260 -23.047 1.00 55.94 N
ATOM 1166 CA MET F 137 -51.070 6.659 -23.479 0.55 59.87 C
ATOM 1168 C MET F 137 -51.180 7.610 -22.289 1.00 60.49 C
ATOM 1169 O MET F 137 -50.426 8.573 -22.173 1.00 61.12 O
ATOM 1170 CB MET F 137 -52.158 7.008 -24.510 0.55 60.96 C
ATOM 1172 CG MET F 137 -52.133 6.227 -25.838 0.55 62.40 C
ATOM 1174 SD MET F 137 -50.694 6.458 -26.932 0.55 72.63 S
ATOM 1176 CE MET F 137 -51.194 5.535 -28.387 0.55 48.25 C

Many thanks to the DisVis Support Team who helped solving the problem very quickly!
Caro

Thank you for sharing!
We will also look into addressing this issue within DisVis so future versions will handle this issue properly!

Best
Joerg